Pavlovian fear conditioning is a ubiquitous form of learning that may underlie disorders of fear and anxiety in humans. Many studies indicate that nuclei in the amygdala including the basolateral complex (BLA) and the central nucleus (CEA) are required for both the acquisition and expression of fear conditioning in mammals. Notably, neurons in both areas exhibit learning-related increases in spike firing during fear conditioning and pharmacological inhibition of both regions reduces fear conditioning. This suggests a role for BLA and CEA neurons in elaborating memories of aversive experiences, although the precise function of these areas in Pavlovian fear conditioning remains unclear. In the present proposal, it is hypothesized that BLA and CEA neurons encode distinct associative representations during fear conditioning. More specifically, we suggest that BLA neurons encode CS-US (S-S) representations that are sensitive to the current value of the US, whereas CEA neurons encode CS-CR (S-R) associations that are insensitive to US value. Moreover, we hypothesize that the relative contribution of these associations, and hence the BLA and CEA, to conditional fear varies as a function of training. To exam these hypotheses, we will use single-unit recordings, reversible pharmacological lesions and sophisticated behavioral designs to isolate the function of amygdala neurons in fear memory processes. The first specific aim will use inflation and devaluation procedures to examine the role of BLA and CEA neurons in representing US "value", which will be indexed by measuring conditional freezing to an auditory CS that has been paired with an aversive US. The second aim will examine the associative basis (S-S or S-R) for conditional fear after overtraining. The third aim will examine whether the role for synaptic plasticity in the BLA and CEA changes as a function of training. Collectively, these experiments represent the first systematic attempt to define parallel, associative functions of BLA and CEA neurons in aversive conditioning. The outcome of these experiments will have an important impact on current conceptions of both amygdaloid function and on clinical disorders of fear that depend on emotional learning and memory.